Listing 1 - 1 of 1 |
Sort by
|
Choose an application
The fire accident in road tunnels can lead to harmful consequence for commuters, as the risk of being killed is higher than open roads. The study of these accidents is limited to some extent. This thesis focuses on multiple risk influencing factors of vehicle collision, scenario construction for understanding tunnel accidents and analysis of large tunnel data using Multiple Correspondence Analysis (MCA). The real-life tunnel accidents are hard to reconstruct due to lack of information gathering and association of different parameters within statistical data. We construct 26 independent accidents with different combination of parameters (tunnel length, tunnel zones, gradient, etc.). The absence of description and graphical representation in tunnel collisions is compensated by assigning a unique data representation for the constructed scenarios. Risk variables or risk factors with multiple categories without any quantitative value is represented as labels. For example, in our thesis the tunnel type is a risk variable which consists of two categories (unidirectional tunnel and bidirectional tunnel). Analysing large amount of nominal data for accident scenarios is done by MCA as it quantifies the categorical value for each representation within an accident. The output obtained is a lower dimensional map of individual scenarios and its categorical variables. This map is directly interpreted for its association from pattern recognition within each scenario. The aim is to show a good data representation for reconstruction of real time accidents and to analyse large categorical crash data for a single tunnel or combination of tunnels.
Listing 1 - 1 of 1 |
Sort by
|